Behavioral and Developmental Homeostasis in the Fire Ant

Journal of Insect Physiology 46 (2000) 933–939 www.elsevier.com/locate/jinsphys

Behavioral and developmental homeostasis in the ﬁre ant, Solenopsis invicta Deby Lee Cassill *, Walter R. Tschinkel Department of Biological Science, Florida State University, Tallahassee, FL 32306-3050, USA

Received 28 May 1999; accepted 5 October 1999

Abstract

The von Bertalanffy rule (1960) predicts that low incubation temperature during larval development will result in larger adult body size. If larval development in social insects followed this rule, then low incubation temperature would induce the development of larger workers and possibly even sexuals. To test this prediction, the effect of incubation temperature on larval development, larval meal size, larval tending and worker recruitment to food in the ﬁre ant, Solenopsis invicta was investigated. Temperatures tested where within the range at which brood remains viable. Contrary to the predictions of the von Bertalanffy rule, worker size was unaffected by incubation temperature, and sexuals were reared at the high rather than the low incubation temperature. Moreover, larval meal size, the rate of larval tending by workers and the total number of workers recruited to food were unaffected by temperature. Mechanisms regulating developmental and behavioral homeostasis were as follows: the duration of larval development and the rate of larval growth changed proportionately with temperature such that the mean and variation of pupal size was unaffected by incubation temperature. Larvae solicited at the same rate, swallowed at the same rate and swallowed for the same duration such that meal size was unaffected by incubation temperature. On the brood pile, fewer workers tended brood at higher incubation temperatures, but worker tempo increased; as a result, brood tending was not adversely affected by incubation temperature. The rate of worker recruitment to food sites outside the nest increased with temperature, but the duration of the recruitment effort decreased such that, over time, the same total number of workers was employed to retrieve food. Incubation humidity was also investigated. When brood chambers were less than 100% humid, workers recruited to food and tended larvae (retrieved, sorted and groomed them), but did not feed larvae. Eventually, larvae died of starvation and were cannibalized.  2000 Elsevier Science Ltd. All rights reserved.

Keywords: Compensatory behavior; Caste determination; Nest; Feeding; Trophallaxis; Worker polyphenism; Formicidae

1. Introduction 1991; Gu and Danthanarayana, 1992; Bouletreau-Merle and Sillans, 1996; Arnett and Gotelli, 1999). Insects are ectotherms, thus their rate of growth, time The relationship between body size and incubation to maturity and ﬁnal adult size may depend upon the temperature for social insects is less well studied. In the temperature of their rearing environment as well as their termite, Coptotermes formosanus, low incubation tem- diet (Slansky and Scriber, 1985; Atkinson, 1994; Atkin- perature produced smaller nymphs (Waller and La Fage, son and Sibly, 1997). The inverse relationship between 1988). In the sweat bee, Halictus ligatus, cool rainy body size and developmental temperature, the von Berta- weather during the breeding season produced smaller lanffy rule (von Bertalanffy, 1960), has been found in body size in workers and alates than warm, dry weather many species of solitary insects (Hogue and Hawkins, (Richards and Packer, 1996). In the ant genus Myrmica, the temperature at which eggs are laid determines whether larvae develop rapidly and become workers or whether they hibernate and retain the potential to * Corresponding author. Tel.: +1-520-621-1151 fax: +1-520-621- become queens (Brian, 1955, 1956, 1963). Hibernating 1150. queen-biased larvae are more likely to succeed in E-mail address: [email protected] (D.L. Cassill). becoming queens when incubated at an intermediate

0022-1910/00/$ - see front matter  2000 Elsevier Science Ltd. All rights reserved. PII: S0022-1910(99)00200-0 934 D.L. Cassill, W.R. Tschinkel / Journal of Insect Physiology 46 (2000) 933–939 temperature than at a low or high temperature (Brian, and maintained at 28°C in constant light for one month. 1973). However, the final size of worker-biased larvae Colonies were fed a diet of tenebrionid beetle larvae and reared early in the breeding season was unaffected by 20% sugar water. For the humidity experiment, year-old incubation temperature (Elmes and Wardlaw, 1983). laboratory-reared colonies were used. These colonies Thus, worker size was independent of incubation tem- were reared from newly-mated, monogyne, Solenopsis perature within the viable range, but adult caste was not. invicta queens collected in Tallahassee, Florida, U.S.A. The fire ant, Solenopsis invicta, is a perennial species during the spring of 1989. Both colony types were main- subjected, not only to diurnal temperature fluctuations, tained by methods similar to those described by Banks et but also to seasonal fluctuations. Nesting in the soil buf- al. (1981). There is no difference in the rates of feeding fers these ants from air temperature extremes. In between laboratory-reared and field-reared larvae addition, the deep and complex nests built by fire ants (Cassill and Tschinkel, 1999b). (Tschinkel, 1998) provide the queen and workers a choice of incubation conditions in which to lay eggs and 2.2. Larval development to rear larvae. Workers move brood up and down temperature and moisture gradients several times a day to To examine the effect of temperature on final adult optimize incubation conditions (Potts et al., 1984; Porter caste or size, larvae from eight source colonies were and Tschinkel, 1993). The incubation temperature for combined, sifted to collect small 4th instar larvae (mesh optimal brood development in this species is 32°C with #35; Cassill and Tschinkel, 1995) and divided into a survival range of 25–35°C (Porter, 1988). Because groups of 0.1 g each (ෂ100 individuals). Each larval thousands of larvae and pupae are distributed among group was placed into an experimental nest containing thousands of small brood chambers at any given time 100 workers. The size of experimental colony fragments (Cassill, 1996), many may experience less than optimal were comparable to the size of groups found in subter- incubation conditions. ranean brood chambers in the field (Cassill, 1996). The If worker body size follows the von Bertalanffy rule final number of brood successfully reared was not (1960), variation in incubation temperature induced by determined. nest complexity could explain the observed variation in Experimental brood chambers were circular with a worker body size within a cohort (Tschinkel, 1988). diameter of 2.25 cm and a depth of 0.3 cm, comparable Moreover, temperature could regulate a beneficial trade- to the mean size of subterranean chambers in the field. off between colony size and worker body size. The Two colony fragments were established from a source observed increase in colony size at higher temperatures colony. One of the pair was incubated at 25°C, the low- (Porter, 1988; Porter and Tschinkel, 1993) might be est temperature at which brood can be successfully mediated by the production of smaller workers (Porter reared. The other was incubated at 32°C, the temperature and Tschinkel, 1986). at which optimal colony growth occurs (Porter and Temperature could regulate individual growth and Tschinkel, 1993). Both were fed 20% sugar water, water development directly via a temperature switch (Brian, and thawed tenebrionid larvae ad libitum. Each week for 1956), indirectly via a nutritional switch (Wheeler, 1986) a total of three weeks, a random sample of 20 white or both (Passera, 1974). If polyphenism was regulated pupae (a surrogate for adult size, Porter and Tschinkel, directly by a temperature switch, cool incubation tempera- 1985) was harvested from each fragment (n=16×3). ture could induce a rise in JH levels, initiating major or Pupae were sampled during the time when both groups gyne development. Subsequently, the rate of larval growth were producing abundant pupae. Pupae were weighed would increase causing larvae to signal more often for individually. Pupae with a mass between 1 mg and 4.25 feedings (Cassill and Tschinkel, 1995, 1999a). Alterna- mg were counted as majors; those with a mass Ͻ1mg tively, temperature-delayed development could result in a were counted as minors. These procedures were repli- larger number of meals for a larva such that, when cated using four different source colonies. nutrition was sufficient, JH levels could rise, initiating major or gyne development. In this paper, we report the 2.3. Larval meal size effects of incubation temperature on larval development, larval meal size, worker brood tending and worker recruit- Larval appetites were quantified using video tech- ment to food. In addition, we report the effects of incu- nology (see Cassill and Tschinkel, 1995, 1999a). Food bation humidity on these same variables. for larval feeding consisted of a distilled water solution with 6% (w/v) casamino acids powder and 10% granu- 2. Materials and methods lated sugar (w/v). We tested temperatures within the range of viable brood production — 25°C, 30°C and 2.1. Source colonies 35°C. From one source colony, 8 g of workers were div- For temperature experiments, mature field colonies ided into three groups of 2 g each. Each group was were collected in Tallahassee during January of 1995 placed into an experimental nest and deprived of food D.L. Cassill, W.R. Tschinkel / Journal of Insect Physiology 46 (2000) 933–939 935 for 48 h. Larvae were collected using the same methods moist nest chambers (100% humidity). A colony was as reported above, then placed temporarily into a single starved for 48 h, then 2 g of workers (ෂ4,000) were ran- brood chamber and deprived of food for 12 h. Nests con- domly aspirated from the brood chamber and foraging taining workers were placed in the temperature-con- arena. Worker groups were halved and each half (1 g) trolled room to acclimate 12 h before the experiment was placed into either a dry brood chamber or a water- began. One gram of larvae was added to each experi- saturated chamber. Next, 1.0 g of larvae were aspirated mental nest 2 h before the experiment began. Fire ants from the source colony’s brood chamber, divided into respond within seconds to the immediate temperature, half and each half (ෂ1,000) was placed in each of two regardless of the historic temperature (Potts et al., 1984). experimental nests. Dyed food (Cassill and Tschinkel, Food was introduced 30 min. before videotaping began. 1995) was introduced into the arena of each experi- Taping lasted 2 h. These procedures were replicated mental nest. After 1 h, the number of fed larvae, as evi- using a total of four source colonies. denced by the presence of dyed food in the mid-gut, was A portion of this experiment was repeated using three counted. Approximately 200 larvae were sampled per additional source colonies tested at 25°C and 35°C. This nest (50 each from four different locations on the brood time, starved larvae were placed in the nest with workers pile). Colony fragments were maintained for three weeks 12 h, instead of 2 h, before food introduction and videot- on a standard diet (see Methods) to determine the fate aping. Temperature probes (VIS Tele-Thermometer; YSI of larvae in dry nests. Replicates from 10 laboratory- 400 series) were placed inside the brood chambers to reared source colonies were completed. confirm that ants were indeed experiencing those temperatures established by the experimental protocol. In 2.7. Data analysis addition, a thermometer was placed vertically on the wall just above the nest and another horizontally on the Dependent variables included: rate of worker-to-larva table beside the nest. Temperatures did not vary more trophallaxis (number/larva/h); duration of trophallaxis ° than 0.5 C among thermometers nor did the temperature (s); rate of larval swallowing (number of boluses/s); dur- ° vary more than 1.0 C during the experiment. Results on ation of larval swallowing (s); lapsed time to first feeding larval feeding from the second experiment did not differ (s); worker density on brood pile (percent of brood pile significantly from the first experiment; therefore, data occupied by workers); rate at which workers encoun- were pooled for analysis. tered a larva (number/h); number of workers recruited to food site; percent of larvae fed; and pupal weight (mg). 2.4. Worker brood tending Depending upon the experiment, data were analyzed using ANOVA or ANCOVA with larval size as a covari- × A grid with 5 5 cm cells was drawn on the monitor ate. ANOVA differences were reported as significant at with a black marker. After pausing the videotape, the PϽ0.05. To determine differences among cell means, cells containing larvae (N) were counted as were the Tukey’s Honestly Significant Difference test was used. cells containing both larvae and workers (n). A cell was Diagnostic analysis was routinely performed to check for considered to contain a worker if at least half of a work- outliers (S.D.Ͼ2.90) and to determine if assumptions of 2 er’s body was within the 25 cm cell. The percent of normality and uniform variance held. Data did not × brood pile coverage was calculated as n/N 100. The vid- require transformation. eotape was paused three times randomly for each treatment in each replicate (n=36). The rate at which workers encountered larvae was determined by counting the number of workers that antennated a larva per hour. 3. Results

2.5. Worker recruitment 3.1. Larval development

For the replicates in which larval feeding was com- No significant difference in mean worker size was pared at 25°C and 35°C, the number of workers recruited found between incubation temperatures (P=0.385) or to the food site was counted after food placement in the among dates (P=0.715). There was a significant differ- arena at 10 min and again at 60 min. Three replicates ence in mean worker size among colonies (PϽ0.001) were completed. with one colony producing larger workers on average than the other three colonies. Interactions among tem- 2.6. Humidity perature, colony and date were not significant. There was no effect of temperature on variance in worker size 2 = = The effect of nest moisture on larval appetite was (COV: X 0.05,1 0.148; P n.s.) nor on worker size within tested on laboratory-reared colonies under two incu- subcaste (major subcaste: P=0.565; minor subcaste: bation conditions, dry (50–65% ambient humidity) or P=0.469). Temperature had no effect on the ratio of 936 D.L. Cassill, W.R. Tschinkel / Journal of Insect Physiology 46 (2000) 933–939 major/minor workers (nonparametric: P=0.118) or on the number of majors produced (X2=3.0, P=0.083). A handful of sexual pupae was produced by each of the colony fragments incubated at 32°C (mean=6 per colony fragment). Sexual pupae were a mixture of males and gynes. No sexuals were produced in colony fragments incubated at 25°C.

3.2. Larval meal size

There was no difference in the rate of worker-to-larva regurgitation among temperatures (P=0.252). Mean rate of regurgitation was 25.0 hϪ1 ±6 S.D.; minimum=5hϪ1; maximum=35 hϪ1. There was no effect of temperature on the lapsed time to the first feeding per larva (P=0.544). Mean lapsed time to the first feeding was 4.0±0.9 min S.D.; minimum=0.5 min; maximum=14.2 min. However, the mean duration of regurgitation was inversely proportional to temperature (PϽ0.001), doub- ling from 11 s at 35°C to 23 s at 25°C. Because low temperature extended the duration of regurgitation, larvae might have been receiving larger meals. Therefore, the rate and duration of larval swallowing was investigated using a dissecting microscope at ×80 magnification and fiber-optic lights. At each temperature, 25°C, 30°Cor35°C, 10 larvae were selected based on their orientation in the brood pile so that swallowing could be observed. Food was dyed green (20% v/v) using across-the-counter food dye to increase the visibility of the bolus. Bolus size does not change with growth in 4th instar (Cassill and Tschinkel, 1996). Temperature did not affect the rate of larval swallowing (2.2±0.2/s, mean±S.E.M.; P=0.914) or the duration of larval swallowing (11.8±1.9 s, mean±S.E.M.; P=0.691). Larvae stopped swallowing after ෂ12 s Fig. 1. Worker brood-tending behavior versus temperature. (a) Per- cent of workers tending larvae per unit of brood pile was inversely regardless of how long workers offered food to them. related to temperature. (b) In contrast, the rate at which workers The drop in the duration of larval swallowing after ෂ10– encountered larvae (number/h) increased proportionately with tempera- 12 s was also observed in hand-fed larvae (Cassill and ture. The end result was that brood tending was unaffected by incu- Tschinkel, 1996). All together, these findings demon- bation temperature. Mean±S.E.M. strated that meal size in larvae was unaffected by incubation temperature. 3.4. Worker recruitment

3.3. Worker brood tending The pattern of recruitment varied signiﬁcantly with time (PϽ0.0001; Fig. 2), and there was a signiﬁcant Worker coverage of the brood pile was inversely pro- interaction between variables (PϽ0.001). Initially, the portional to temperature, decreasing about 30% as tem- number of workers recruited to food at 35°C was twice perature increased from 25°Cto35°C(PϽ0.003; Fig. the number of those recruited at 25°C. After 60 min, the 1a). Despite the decrease in the number of workers tend- reverse was true. Thus, nearly the same total number ing brood, the rate of larval encounter by workers of workers was recruited to the food site regardless of increased almost two-fold with temperature (PϽ0.001; temperature (P=0.212). Fig. 1b). To attain this increase in larval contact with fewer workers, the tempo of workers had to increase 3.5. Humidity three-fold over the range of temperatures from 25°Cto 35°C — a 30% increase in worker tempo per degree Four hours after food introduction, all larvae housed increase in temperature. in moist chambers (100% humidity) were fed while none D.L. Cassill, W.R. Tschinkel / Journal of Insect Physiology 46 (2000) 933–939 937

can be inferred from the size of workers at eclosion. If larval growth rate is temperature-independent, then temperature-delayed development in a cool incubation chamber should result in a larger pupa. This hypothesis is consistent with the von Bertalanffy rule (1960). How- ever, if larval growth rate is temperature-dependent, then larvae will pupate at the same size regardless of incubation temperature. We found that worker-biased larvae pupated at the same mean size regardless of incubation temperature. Thus, developmental homeostasis is achieved by a temperature-dependent growth rate as well as a temperature-dependent developmental duration. Both changed similarly with temperature. Worker size is an important factor in division of labor (Cassill and Tschinkel, 1999c) and worker longevity (Calabi and Porter, 1989) in the fire ant. Therefore, a stable mean and a stable variation in worker body size Fig. 2. Workers recruitment to food versus temperature. When ambient temperature was 35°C, the initial rate of recruitment was high, but may be an adaptation by this perennial species to cope declined quickly. In contrast, when ambient temperature was 25°C, the with diurnal and seasonal fluctuations in temperature. In initial rate of recruitment was low but continued at that level for an support of this speculation is the fact that mean worker extended period. In the end, the same total number of workers was size does not vary within or among seasons in mature recruited to the food site regardless of temperature. fire ant colonies (unpublished data). The fire ant was an exception to the von Bertalanffy housed in dry chambers (ෂ50–65% humidity) was fed. rule in a second way. Sexuals (males and gynes), which Larvae were far more densely piled in dry chambers, are far larger than workers, were produced at high rather taking up about one-half the surface area of those in than low incubation temperature. The presence of males dampened chambers. Over the next two weeks, regard- in one but not the other treatment suggests that workers less of nest moisture, workers recruited to food, shared cannibalized sexual larvae at the low temperature. Both food with other workers via regurgitation, and groomed of our findings, temperature dependent sexual production larvae (presence/absence of these behaviors was and temperature independent worker body size, were recorded). However, even in the presence of abundant consistent with another ant species, Myrmica (Elmes and food, larvae in the dry chambers starved to death and Wardlaw, 1983). Because we did not test the effects of eventually were cannibalized. temperature on eggs or early instar larvae, we can not rule out the possibility that worker body size is temperature sensitive, if exposure occurs at an earlier develop- 4. Discussion mental stage. Until additional experiments are completed, the issues of whether caste determination and Contrary to the predictions of the von Bertalanffy rule body size are regulated by the same or different mech- (1960), neither worker body size nor larval meal size anisms, and whether these mechanisms include a tem- was affected by incubation temperature. Homeostasis in perature switch (Brian, 1956, 1973), a nutritional switch the development of worker-biased larvae and in their (Wheeler, 1986, 1990) or both (Passera, 1974), remain meal size was achieved by a combination of compensa- unanswered. tory development and behaviors. Additionally, sexuals, which are far larger in body size than workers, were 4.2. Behavioral homeostasis reared only at high rather than the low incubation temperature. A discussion of the compensatory mechanisms Despite a 10°C difference in the temperature within regulating the homeostatic development and behavior of brood chambers, larvae received the same total volume worker-biased larvae follows. of food. Meal size homeostasis was achieved through several compensatory mechanisms. First, although tem- 4.1. Developmental homeostasis perature extended the duration of each worker-to-larva regurgitation event, the rate and duration of larval swal- Although temperature can extend the duration of lar- lowing was unaffected. Thus, larvae ingested the same val development two-fold (Porter, 1988; Porter and volume of food regardless of how long a worker Tschinkel, 1993), little is known of its effect on the rate presented it to them. Second, worker number and worker of larval growth. There are two possible relationships tempo varied with temperature in such a way that the between temperature and growth rate, either of which rate of brood tending by workers was homeostatic. 938 D.L. Cassill, W.R. Tschinkel / Journal of Insect Physiology 46 (2000) 933–939

Third, colonies achieved recruitment homeostasis by Atkinson, D., 1994. Temperature and organism size — a biological increasing the initial rate of recruitment, but decreasing law for ectotherms? Adv. Ecol. Res. 25, 1–58. Atkinson, D., Sibly, R.M., 1997. Why are organisms usually bigger in its duration at high temperature relative to low tempera- colder environments? Making sense of a life history puzzle. Trends ture (Fig. 2). The mechanism by which colonies modu- Evol. Ecol. 12, 235–239. late recruitment rate and duration at different tempera- Banks, W.A., Lofgren, C.S., Jouvenez, D.P., Stringer, C.E., Bishop, tures is unknown. P.M., Williams, D.F., Wojcik, D.P., Glancey, B.M., 1981. Tech- niques for rearing, collecting and handling imported fire ants. USDA and SEA Agriculture and Technology Southern Service 21, 4.3. Humidity 1–9. Bouletreau-Merle, J., Sillans, D., 1996. Effects of interaction between Although the effect of humidity on worker mortality is temperature and CO2 on life-history traits of two Drosophila spec- minimal (Hood and Tschinkel, 1990), its effect on larval ies (Diptera: Drosophilidae). Eur. J. Entomol. 93, 451–459. Brian, M.V., 1955. Studies of caste differentiation in Myrmia rubra mortality was dramatic. In dry nests, larvae were never L. 3. Larval dormancy, winter size and vernalisation. Insectes Soc- fed even though workers were able to retrieve, sort, pile iaux 2, 85–114. and groom them. The end result was that larvae died and Brian, M.V., 1956. Studies of caste differentiation in Myrmia rubra were eventually cannibalized before they could pupate. L. 4. Controlled larval nutrition. Insectes Sociaux 3, 369–394. We speculate that the mechanism by which larvae reduce Brian, M.V., 1963. Studies of caste differentiation in Myrmica rubra L. 6. Factors influencing the course of female development in the their rate of desiccation is linked to their ability to signal early third instar. Insectes Sociaux 19, 91–102. for feedings. Brian, M.V., 1973. Temperature choice and its relevance to brood sur- In summary, our findings demonstrated that worker- vival and caste determination in the ant Myrmica rubra L. Physiol. biased larvae achieved developmental homeostasis in the Zool. 46, 245–252. face of variable incubation temperature through com- Calabi, P., Porter, S.D., 1989. Worker longevity in the fire ant Solen- opsis invicta: ergonomic considerations of correlations between pensatory behaviors and development. Such homeostasis temperature, size and metabolic rates. J. Insect Physiol. 35, 643– may be an adaptation by the fire ant to its perennial and 649. weedy life style. Hogue and Hawkins (1991) found that Cassill, D.L., 1996. Flow of food and social organization in the fire temperature affected body size in species with a simple ant, Solenopsis invicta, The Florida State University, Tallahassee, life-history pattern, but not in species with more com- Florida, p. 307. Cassill, D.L., Tschinkel, W.R., 1995. Allocation of liquid food to lar- plex life histories. Moreover, our results suggest that vae via trophallaxis in colonies of fire ants, Solenopsis invicta. Ani- caste determination and body size may be regulated by mal Behav. 50, 801–813. entirely different genetic mechanisms or environmental Cassill, D.L., Tschinkel, W.R., 1996. A duration constant for worker- triggers. In the final analysis, although much is known larva trophallaxis in ants. Insectes Sociaux 43, 149–166. about the role of juvenile hormone in the timing of the Cassill, D.L., Tschinkel, W.R., 1999a. Regulation of diet in the fire ant, Solenopsis invicta. J. Insect Behav. 12, 307–328. developmental switch from one caste to another in ants Cassill, D.L., Tschinkel, W.R., 1999b. Effects of colony-level attri- (Wheeler, 1986, 1990, 1991, 1994), there remains much butes on larval feeding in the fire ant, Solenopsis invicta. Insectes to be learned about the internal and external regulation Sociaux 26, 261–266. of that switch. Cassill, D.L., Tschinkel, W.R., 1999c. Task selection by workers of the fire ant, Solenopsis invicta. Behav. Ecol. Sociobiol. 45, 301–310. Elmes, G.W., Wardlaw, J.C., 1983. A comparison of the effect of temperature on the development of large hibernated larvae of four Acknowledgements species of Myrmica (Hym Formicidae). Insectes Sociaux 30, 106–118. We thank Michael Sloderbeck for programming the Gu, H., Danthanarayana, W., 1992. Influence of larval rearing conditions on the body size and flight capacity of epiphyas-postvittana computerized event recorder “Timeant3” and Nick Dai- moths. Aust. J. Zool. 40, 573–581. gle for transcribing video-taped feeding interactions. Hogue, J.N., Hawkins, C.P., 1991. Developmental temperature and Additionally, we thank Diana Wheeler and two anony- seasonal growth-patterns. J. North Am. Benthol. Soc. 10, 309–321. mous referees for critical reviews of the manuscript. This Hood, W.G., Tschinkel, W.R., 1990. Desiccation resistance in arboreal project was completed with support from the Center for and terrestrial ants. Physiol. Entomol. 15, 23–35. Passera, P.L., 1974. Differenciation des soldats chez la fourmi Phei- Insect Science’s National Institute of Health grant I-T32- dole pallidula NYL, (Formicidae Myrmicinae). Insectes Socieaux A107475 and the National Science Foundation grant 21, 71–86. IBN-9317853. Porter, S.D., 1988. Impact of temperature on colony growth and developmental rates of the ant, Solenopsis invicta. J. Insect Physiol. 34, 1127–1133. Porter, S.D., Tschinkel, W.R., 1985. Fire ant polymophism References (Hymentoptera: Formicidae): factors affecting worker size. Ann. Entomol. Soc. Am. 78, 381–386. Arnett, A.E., Gotelli, N.J., 1999. Bergmann’s rule in the ant lion Myr- Porter, S.D., Tschinkel, W.R., 1986. Adaptive value of nanitic workers meleon immaculatus DeGeer (Neuroptera: Myrmeleontidae): geo- in newly founded red imported fire ant colonies (Hymenoptera: graphic variation in body size and heterozygosity. J. Biogeog. 26, Formicidae). Ann. Entomol. Soc. Am. 79, 723–726. 275–283. Porter, S.D., Tschinkel, W.R., 1993. Fire ant thermal preferences: D.L. Cassill, W.R. Tschinkel / Journal of Insect Physiology 46 (2000) 933–939 939

behavioral control of growth and metabolism. Behav. Ecol. von Bertalanffy, L., 1960. Principles and theory of growth. In: Nowin- Sociobiol. 32, 321–329. ski, W.N. (Ed.) Fundamental Aspects of Normal and Malignant Potts, L.R., Francke, O.F., Cokendolpher, J.C., 1984. Humidity prefer- Growth. Elsevier Press, Amsterdam, pp. 137–259. ences of four species of fire ants (Hymenoptera: Formicidae: Waller, D.A., La Fage, J.P., 1988. Environmental influence on soldier Solenopsis). Insectes Sociaux 31, 335–339. differentiation in Coptotermes formosanus shiraki Richards, M.H., Packer, L., 1996. The socioecology of body size vari- (Rhinotermitidae). Insectes Sociaux 35, 144–152. ation in the primitively eusocial sweat bee, Halictus ligatus Wheeler, D.E., 1986. Developmental and physiological determinants (Hymenoptera: Halictidae). OIKOS 77, 68–76. of caste in social Hymenoptera: evolutionary implications. Am. Slansky, F. Jr., Scriber, J.M., 1985. Food consumption and utilization. Naturalist 128, 13–34. In:. Kerbut, G.A., Gilbert, L.A. (Eds.), Comprehensive Insect Wheeler, D.E., 1990. The developmental basis of worker polymor- Physiology, Biochemistry and Pharmacology, vol. 14. Pergamon phism in fire ants. J. Insect Physiol. 36, 315–322. Press, Oxford, pp. 87–163. Wheeler, D.E., 1991. The developmental basis of worker caste poly- Tschinkel, W.R., 1988. Colony growth and the ontogeny of worker morphism in ants. Am. Naturalist 138, 1218–1238. polymorphism in the fire ant, Solenopsis invicta. Behav. Ecol. Wheeler, D.E., 1994. Nourishment in ants: patterns in individuals and Sociobiol. 22, 103–115. societies. In: Hunt, J., Nalepa, C.A. (Eds.) Nourishment and Evol- Tschinkel, W.R., 1998. The reproductive biology of fire ant societies. ution in Insect Societies. Westview Press, Boulder, Colorado, pp. Biol. Sci. 48, 593–605. 245–278.